Pixel binning
One reason to do pixel binning is to increase the SNR. Almost the same signal is falling on adjacent pixels, especially if they are small, but the noise is random so it's different for adjacent pixels.
The Lumia 1020, released in 2013, has a 41 megapixel sensor. The RAW image is 7136 x 5360 and the standard JPEG is 2592 x 1936 (both saved at the same time). I believe it is using pixel binning. And either the Sony a7RII or the a7SII (can't remember which) does pixel binning in HD video mode, the other selected 1920x1080 pixels evenly spaced across the sensor. The difference is noticeable in low light as low light which drives down the SNR.
However, the 1020 illustrates other issue. It had a larger than normal lens made by Zeiss. The true/useful resolution, measured objectively with a camera chart, is the least of the sensor megapixels and the resolving power of the lens. Many lenses even on DSLR cameras max out before the sensor does and that is much more likely to be the case on a phone camera. Many cinematography lenses in use max out around 4k horizontal resolution. That's not a problem on an ARRI Alexa or 35mm film where the resolution is <4k, but if you use them on an 8k RED Mostro you're not going to see a 2x increase in sharpness.